DEPARTMENT OF CHEMICAL AND METALLURGICAL ENGINEERING Heat Transfer Laboratory The University of Michigan Ann Arbor, Michigan Progress Report on THE PERFORMANCE OF HIGH-FIN TUBES IN COMBINED RADIATIVE AND NATURAL CONVECTION HEAT TRANSFER Report No. 57 Edwin H. Young Professor of Chemical and Metallurgical Engineering Charles W. Bash Madhukar V. Kulkarni David H. Wierenga Research Assistants Project 1592 WOLVERINE TUBE Division of CALUMET &HECLA, INCORPORATED ALLEN PARK, MICHIGAN January 1965

TABLE OF CONTENTS List of Tables......................... List of Figures......................... Notation............................... 0 0 * 4 0 * 0 0. 0 f. 0 * * 6. * 0 0 * 0 0 0 aa a 0 0. 0 * 0 0 0* 4 0 Page ii iii v Abstract................................ Objective......... *................ Introduction............................. Selection of Experimental Heat Exchanger.. Survey of the Technical Literature.......... Laboratory Apparatus..................... Operation of Equipment.................... Stack Gas and Fuel Gas Analysis........... Analysis of Data.......................... Heat Balance...................... Radiation.......................... Example Analysis.................. Water Film Coefficient............. Gas Film Coefficient............... 0.0 0 0. v 0 a 9 0. 0 0 * - 0* a 0 a * 0 0. 0 0 0 0 * 0 0 0 * 0 0 0 0 0. 0 * 0 0 & 9. a.. *.. 0... 0 0 0 00.* * 0 0 0 0 0 4 0 0 & 0 0 4 * * 0 1 16...... 2 27...... 22......27...... 30...... 33.....33 ~~~~~~~~~~~~~~~~~ ~~CI~~~~O~~C~~~~~ ~~~~~~~~~O~~~Q~~~ ~~~~~~~~~~~~~~~~~ ~~~~~~~~~0~~~~~~~ ~~~~~~~~~~~~~~~~~ ~~~~~~~~~*~~~~~~~ ~~~~~~~~~~~0~~~~* 0 0 0 &. 34 Discussion of Results.......................................... Modifications to the Experimental Apparatus and Future Work........ Literature Cited............................................... Appendix......................................... a. Calculation of Mass Spectrometer Results................... Physical Properties of Carbon Dioxide................... Nitrogen........................... Oxygen......................... Water Vapor....................... 37 44 46 50 51 56 59 62 65 i

LIST OF TABLES Tables Page I Catalogue Number of Tubes Purchased for Hot Water Heater Units. 2 II Wolverine Finned Tubes Purchased by Various Hot Water Heater Manufacturers.... s........o........................... 3 III Details of Finned Tubes Supplied with the Water Heater.......... 11 IV Hot Gas to Fin Tube Heat Transfer........................ 18 V Data Sheet No. 2.................................... 19 VI Data Sheet No. 3.............*................... 20 VII Data Sheet No. 4.......................................... 21 VIII Comparison of Fuel Gas Compositions......................... 25 IX Comparison of Flue Gas Analyses............................,. 26 X Analysis of Data Collected for Experimental Run No. 7...... 30 XI Revised Heat Balance for Run No. 7............................ 32 XII Calculated Coefficients and Resistances............,......,. 36 XIII References Used for Physical Properties of Flue Gas........... 38 XIV Tubes on Order for Investigation..................,........ 44 XV Typical Digital Computer Output for Mass Spectrometer Analysis of Natural Gas..........,.................. 53 XVI Typical Digital Computer Output for Mass Spectrometer Analysis of Stack Gas for Run No. 7A........................ 54 XVII Typical Digital Computer Output for Mass Spectrometer Analysis of Stack Gas for Run No. 7C....................... 55 ii

LIST OF FIGURES Figure Page 1 Laboratory Experimental Installation.................... 6 2 Manufacturers Rating Curve of Recovery Rate. Input Rate is 2,400,000 Btu/hr.............................. 7 3 Line Diagram of Experimental Apparatus.................. 12 4 Laboratory Experimental Installation.................... 14 5 Mass Spectrometer Apparatus.......................... 23 6 Typical Mass Spectrometer Plot........................ 24 7 Rate of Luminous Radiation Heat Transfer............... 29 8 Non-luminous Gas Radiation Heat Transfer Coefficient.... 31 9 Theoretical Temperature of Gas Leaving Burners......... 39 10 Dewpoint of Flue Gas.................................. 41 11 Gas Film Coefficient for Triangular Pitch Banks......... 42 12 Sketch of Thermocouple Shield Assembly................ 45 13 Thermal Conductivity of Carbon Dioxide................ 56 14 Heat Capacity of Carbon Dioxide........................ 57 15 Viscosity of Carbon Dioxide........................... 58 16 Thermal Conductivity of Nitrogen....................... 59 17 Heat Capacity of Nitrogen.............................. 60 18 Viscosity of Nitrogen.................................. 61 19 Thermal Conductivity of Oxygen................... 62 20 Heat Capacity of Oxygen............................... 63 21 Viscosity of Oxygen................................... 64 iii

LIST OF FIGURES (Continued) Figure Page 22 Thermal Conductivity of Water Vapor................... 65 23 Heat Capacity of Water Vapor........................... 66 24 Viscosity of Water Vapor............................... 67 iv

NOTATION A = Mean area in sq.ft./ft. m A = Area, sq. ft. A = Area of fins per foot of length, sq.ft. /ft. f A. = Inside heat transfer area per foot of length, sq.ft. /ft. 1 A = Outside heat transfer area per foot of length, sq.ft. /ft. O A = Area of root per foot of length, sq.ft. /ft. r D. = Inside diameter of tube, inches D = Outside diameter of bare tube or diameter over the fins of finned o' tubes, inches D = Root diameter of tube, inches r F = Radiation interchange shape factor from a to b, dimensionless ab G = Mass flow rate at minimum cross section, lb. /hr.-sq.ft. max H = Fin height, in. h. = Mean inside heat transfer coefficient, Btu/hr.-sq.ft.(internal area) ~F 1 h = Mean outside heat transfer coefficient, Btu/hr-sq.ft. (external area) ~F 0 h = Radiative heat transfer coefficient Btu/hr.-sq.ft.(root area)~R r k = Thermal conductivity of gases, Btu/hr.-sq.ft.- ~F/ft. k = Thermal conductivity of fin material, Btu/hr.-sq.ft. - ~F/ft. m I = Fin height, inches N = Number of fins per linear inch Pr = Prandtl Number, Cpp/k Q = Heat duty, Btu/hr. Re = Reynolds Number,(D /G )/I r max rf = Fin resistance, hr.- ~F-sq.ft./Btu rf v

r = Metal resistance, hr. - ~F-sq.ft. /Btu m S = Distance between adjacent fins, in. t = Fin thickness, inch t = Mean water temperature, ~F w T = Absolute temperature, ~R U =Overall heat transfer coefficient based on outside area, ~0 Btu/hr.-sq.ft.- ~F V = Velocity of water in tube, ft./sec. t y = Fin thickness, ft. AT =Logarithmic temperature difference, ~F m - = Viscosity at bulk stream temperature, lb. /ft.-hr. -9 o' = Radiation constant, 1.712 x 10 Btu/hr.-sq.ft.- ~R E Emissivity of surface i, dimensionless vi

ABSTRACT The summary of the status of the laboratory investigation currently in progress on the performance of high-fin tubes in combined radiative and convective heat transfer in direct fired heating applications is presented. Insufficient experimental data has been obtained to date on enough banks of finned tubes containing various sizes of fin heights, fin spacings, fin thickness, and root diameter with various tube pitch arrangements to make any predictions or draw any significant conclusions at this time. OBJECTIVE The purpose of this investigation is to obtain experimental heat transfer and pressure drop data on high-fin tubes in combined radiative and natural convection heat transfer using hot flue gas from a direct-fired heater. The experimental data is to be analyzed and correlated using regression analysis techniques on the IBM 7090 digital computer. 1

INTRODUCTION Wolverine Tube has been providing a number of hot water heater manufacturers finned tubes for use in their commercial units. Table I presents the catalogue number of the tubes purchased from Wolverine Tube by eight hot water heater manufacturing companies. TABLE I Catalogue Number of Tubes Purchased for Hot Water Heater Units Ace Tank & Heater Raypak Company 61-0516065-01 61-0710049-01 61-0520065-01 61-0714065-01 61=0714072-01 Rheem Manufacturing Amarillo Welding 61-0516065-01 63-0710065-01 63-0712065-01 Swimming Pool Supply 61-0710065-01 Fleetwood Mfg. Company 610714072-01 61-0714072-01 61-0710049-01 61-0710065-01 Weben Industries e 6 1 05 61-05 16065-01 Laars Engineering 61-0614065-01 61-0710065-01 61-0714072-01 An examination of Table I indicates that hot water heater manufacturers purchasing tubes have been purchasing tubes having a wide range of dimensions for the same heat transfer application. In several instances, some of the manufacturers purchase the same tube. Table II indicates which manufacturers purchase which tube. Nine different tubes are purchased. An examination of the catalogue dimensions corresponding to the tubes presented in Table II indicates that hot water heater manufacturers are purchasing finned tubes having 5 to 7 fins per inch, nominal inside diameters varying from 5/8 in. to 1 1/4 in., root wall thicknesses varying from 0.049 in. to 0.072 in., fin diameters varying from 1.4 in. to 2s2 in,, fin heights from 0.35 in. to 0.40 in. All of the tubes have the same mean fin thickness. Consequently, the outside 2

TABLE II Wolverine Finned Tubes Purchased by Various Hot Water Heater Manufacturers Tube Catalogue No. 61-0516065-01 61-0520065-01 61-0614065-01 61-0710049-01 61-0710065-01 61-0714065-01 61-0714072-01 63-0710065-01 63-0712065-01 Organization Ace Tank & Heater Rheem Manufacturing Weben Industries Ace Tank & Heater Laars Engineering Fleetwood Mfg. Company Raypak Company Fleetwood Mfg. Company Laars Engineering Swimming Pool Supply Raypak Company Ace Tank & Heater Laars Engineering Swimming Pool Supply Amarillo Welding Amarillo Welding areas and the inside areas in square feet per foot of tube length vary over a wide range, as do also the outside to inside surface area ratios. There appears to be no logical reason why the eight manufacturers involved should purchase finned tubes for the same application that have such a wide variation in dimensional char — acteristics o The purpose of this investigation is to explore the effect of a wide range of dimensional characteristics on the performance of high fin tubes in this type of heat transfer application. The results obtained from this investigation should enable one to design an optimum tube for this application. 3

SELECTION OF EXPERIMENTAL HEAT EXCHANGER Initially, it was felt that in order to obtain the quality of experimental heat transfer data needed for this investigation, it would be necessary to design a special natural gas fired or liquid petroleum gas (LPG) fired heat exchanger and have it fabricated under contract to some shop. Before proceeding with the development of such a design, catalogues were obtained from all of the significant hot water heater manufacturers. The package units available were carefully evaluated and it was found that several possibilities existed. Discussions with The University of Michigan shop personnel indicated that neither they nor any shop could fabricate a special unit for anywhere near the cost of a commercially built package unit. It was decided that, for economic reasons, a package unit should be used. An evaluation of LPG vs. natural gas indicated that none of the LPG manufacturers control the composition of their fuel close enough for the needs of the research project. In fact, the variation in composition from load to load is quite significant. This would result in widely fluctuating heating values and associated problems. Because of the quantity of LPG that would be needed in storage and far hourly consumption, a vaporization unit would be required. For fire control reasons, the storage tank and vaporizer would have to be installed at a minimum of 75 feet from the Fluids Building. It turned out that The University would require that this unit be installed across the road in the woods behind the Fluids Building. For these reasons and because the pressure of the vaporized gas at the inlet to the flow regulators and meters adjacent to the experimental unit would fluctuate too widely, LPG was, therefore, discarded as the heating medium. Discussions with representatives of the Michigan Consolidated Gas Company indicated that they control the heating value of natural gas delivered to the gas main at + 2%. Also, the pressure on the gas main serving the'North Campus buildings is maintained fairly constant. The pressure reducing regulation system installed by the gas company in the Fluids Building drops the gas main pressure to a carefully regulated pressure inside the building. A further drop in pressure would be required at our experimental unit. This could be satisfactorily controlled without any difficulty. In addition, the gas company offered to loan the project group a Rockwell Model No. 4 gas meter which has a capacity of 2,250 cu. ft. per hour at 1/2 inch water pressure and 5,000 cu. ft. per hour at 1 inch water pressure. The meter is worth approximately $2,800. All factors involved clearly indicated that natural gas should be used as the fuel. An evaluation of the various types of gas burners was made. The uniformity of the burner gas temperature is a significant variable in the experimental investigation. One type of burner appeared to be far superior to all others being used in this respect. This burner is referred to as a "multi-blade" or "strip" burner. Two manufacturers use this type of burner, Weben Industries, Inc. and Raypak: Company, Inc. Discussions with these two manufacturers indicated that only the Raypak Company could provide a large enough unit with enough flexibility for the 4

investigagion. A Raypak Coppertherm Series II unit with modified instrumentation having a maximum input rating of 2,400,000 BTU/hour was purchased and installed on the ground floor of the Fluids Building located at the North Campus of The University of Michigano Figure 1 shows the installed unit. Figure 2 presents the recovery rate curve for the unit as provided by the manufacturer. The Fluids Building has a cooling tower water loop of sufficient capacity to handle the water flow rates required for the investigation. Sufficient natural gas capacity is available in the building to supply the burner requirements of the unit. 5

0 +4 (r p-4 0W.~ >1 mai 0f 6

300 200 150 - 100 - 0 c \ 50k 30 - 20 15 10 1,000 2,000 5,000 10,000 20,000 409000 Water flow rate, gal./min. Fig. 2 Manufacturers Rating Curve of Recovery Rate. Input Rate is 2,400,000 Btu/hr. 7

SURVEY OF THE TECHNICAL LITERATURE The American Gas Association has made many studies of the use of natural gas for hot water production. Most of these are concerned with domestic heaters with the hot water tank being part of the flue (1, 2, 3, 4) and with burner design for this type of service (5, 6, 7). Studies have also been made of gas fired space heaters (8), and application to simultaneous space heating and hot water production (9). The American Gas Association has made a very extensive study of condensation from natural gas (10). In this study they investigated the dewpoint of the flue gas as a function of excess air and sulfide content in the fuel. It was found that 10 grains of hydrogen sulfide per 100 cu. ft. of fuel gas raised the dew point of a typical flue gas (7.2%0 CO ) from 125 ~F to 140 ~F. It was also found that condensation improved the rate of heat transfer by approximately 12%o. A survey of the available technical literature on the subject of combined radiation and convective heat transfer to fin-tubes indicates that most of these papers are for fin tubes in radiant sections and convective sections in fired heaters used in oil refineries and gasoline plants. The most recent useful technical paper is that of Schweppe and ToQrTxjos(ll). This paper was published in June of 1964 and concerns the increase in heat recovery of refinery heating equipment by using fintubes in the convective section of the heater. The procedures presented are based on a master's thesis by Torijos (12) and on methods presented by Lobo and Evans (13) for combustion section radiation; Monrad (14) and Briggs and Young (15) for heat transfer within the convective. section; Gardner (16) for efficiency of extended surfaces; and Gunter and Shaw (17) for pressure drop in crossflow. In 1963, Wimpress (18) of the C. F. Braun and Company published a special report on rating fired heaters. This particular article presents a method which can be used to predict the performance of direct-fired heaters in typical refinery and petrochemical process heating applications. The method for rating the radiant sections is based on an earlier paper by Wilson, Lobo and Hottel (19). Mieth presented a paper entitled "The Importance of Fin-Tube Application in Energy Conservation" (20) before the Petroleum Mechanical Engineering Conference of ASME in September of 1964. This article presents some test data for finned tubes of various types and bare tubes in waste-heat recovery applications. The American Standards Association has approved a set of approval requirements for gas water heaters (21, 22, 23, 24). These standards have primarily to do with safety aspects of the design, construction and installation of such units. The American Gas Association's approval program for gas appliances is based upon American Standards, and the American Gas Association sponsors the American Standards Association project in this field. 8

The technical literature was reviewed to determine what flue gas temperature measurement procedures are acceptable. A symposium on flue gas temperature measurement was held at the American Society of Heating, Refrigeration and Air Conditioning Engineers Semi-Annual Meeting in New Orleans in January 1964. At this symposium, W. B. Kirk, Chief Research Engineer, American Gas Association Laboratories, Inc., presented a comparison of the procedures under the American Standard Association approval requirements for gas appliances (25). K. 0, Schlentner, Director of Engineering, Crane Company, presented flue gas temperature measurement procedures followed by the Institute of Boiler and Radiator Manufacturers and the Steel Boiler Institute (26). R. E. Barrett and H. R. Hazard of the Battelle Memorial Institute presented a paper on the problems in fluegas temperature measurement (27). They examined the codes of the Institute of Boiler and Radiator Manufacturers, American Standards Association-American Gas Association, American Society of Heating and Air Conditioning Engineers (predecessor to ASHRAE), Boiler Manufacturers Association, Steel Boiler Institute, Underwriters Laboratories, and Commercial Standards and found that recommended procedures differ significantly, so that different values of gas temperatures result from the use of these codes. They concluded that accuracy and reproducibility of flue-gas temperature measurements would be improved by codes incorporating the following provisions: (1) insulated flue pipes; (2) measurements made at a uniform location, close to the furnace; (3) smaller thermocouple wires and beads; and (4) multiple-thermocouple grids. In general, there are two standard techniques, one known as the "5-point thermocouple method" and the other known as the "orifice method." The Institute of Boiler and Radiator Manufacturers and the Steel Boiler Institute Codes accept the American Gas Association 5-point thermocouple method of stack temperature measurement as an alternate to the orifice method. The 5-point thermocouple method results in good correlation with the orifice method. A careful study of the two methods indicated that the 5-point thermocouple method was well suited to our laboratory heater installation without major modification. It appeared quite impractical to use the orifice method. Consequently, the 5-point thermocouple method was adopted and the installation of the required thermocouples made. In general, only the bottom horizontal row of finned tubes will be greatly affected by radiation when rectangular banks of finned tubes are used. The second horizontal row of finned tubes will be affected when triangular pitch banks of finned tubes are used. Insofar as convective heat transfer alone is concerned, generalized correlations only for triangular pitch banks of finned tubes are available. None exist for square pitch or rectangular pitch finned tube banks. The most useful generalized correlation for triangular pitch banks of finned tubes is that obtained by the Wolverine Tube Project at The University of Michigan. The initial laboratory work was carried out by Dennis J. Ward on his doctoral thesis under a Wolverine Tube Fellowship (28). This work was later extended by the Wolverine Tube Project group and a correlation, including the effect of fin thickness, fin height, and fin spacing,was published in the technical literature,by:.Briggs and Young (15). The latter reference will be of considerable value in this investigation. 9

Two books on the general subject of combustion contain some useful information. The books are Combustion Engineering" by deLorenzi (29) and the "North American Combustion Handbook" (30). A thorough search was made of the literature in an attempt to find the viscosity, thermal conductivity and heat capacity of the flue gas components for a temperature range of 500 ~F to 2,000 ~F. The major source was National Bureau of Standards (NBS) Circular 564 (31). This source is complete only for viscosity data. Heat capacity data was obtained for all components and checked against the NBS Circular from McBride's listing of thermodynamic properties (32). Thermal conductivity data was reported at the Fourth World Power Conference (33). This data does not agree with that of Keenan and Keyes (34) for water vapor. The Fourth World Power Conference data did agree with the NBS Circular data in the case of oxygen and carbon dioxide and were therefore used to extend the curves of oxygen and carbon dioxide. Thermal conductivity data is also reported for carbon dioxide by Geringer (35). 10

LABORATORY APPARATUS The experimental equipment is basically the Raypak boiler "Coppertherm", Series II Model 2400. This unit has an input rating of 2,400,000 Btu/hr and an output rating of 1,920,000 Btu/hr. As supplied, the water flows through four passes in a bank of 24 tubes (No. 61-0714065-01). Table III lists the physical dimensions and properties of these tubes. The tubes are placed in groups of 12 tubes, two rows deep. Vertical spacing varies along the length of the tubes, horizontal spacing is with touching fins. The center area is blocked off with a fabricated steel strip. TABLE III Details of Finned Tubes Supplied with the Water Heater K, thermal conductivity of fin material m D, outside fin diameter, inches o D, root diameter, inches r Do, inside diameter, inches 1 S, fin spacing, inches Fins /inch t, fin thickness, inches 196 Btu/hr. sq.ft. / ~F/ft. =1.7914 1.000 =0.875 0.1218 = 7 = 0.021 H, fin height A, outside area, sq.ft. /lin.ft. o A,, inside area, sq.fto/lin.ft. 1 A /A, o 1 D /D. 0 1 Af, area of fin, sq.ft. /ft. A, area of root, sq.ft. /ft. r A /A r f D /D o r y, fin thickness, fto = 0.3957 2.316 =0.229 =10.1 1.7914 =2.093 0.223 0.1065 1.223 = 0.00175 A line diagram of the experimental layout is shown in Figure 3. Gas is fed to six burners; each is equipped with an automatic modulation valve, which are set to maximum to prevent interference with data collection. The six burners are fed from two manifolds controlled by electrical cut-off valves. These valves are connected such that if (a) any of the four pilot lights are off, (b) if there is no water 11

TC 150 # STEAM -O GAS ANALYSIS TRC I X 4" _ THERMOMETER & THERMO TO SUMP IMOMETER TM WATER Fig. 3 Line Diagram of Experimental Apparatus

flow, or (c) if the main burner does not produce flame within eight seconds of being turned on, they will automatically close. A further interlock provides that if the temperature of the exit water goes above a specified limit (210 ~F), the gas supply will be automatically cut off. The two gas headers are supplied from the gas main after a meter, a pressure regulator (set af eight inches of water), a master cut-off valve (manual), and a throttling valve. The throttling valve is controlled by a Metrol constant pressure controller so as to maintain constant pressure in the headers. This pressure can be varied from 0.86 to 6 inches of water. A Metrol meter is connected to monitor the header pressure in the range 0 to 4 inches of water. Water is supplied from a cooling tower provided at the Fluids Building. Valves are provided to completely isolate the cappa-iatus. from the cooling tower and provision is made to drain the system of water, which is done after every experimental run. The inlet water flows through an orifice and then a valve controlled by a Minneapolis -Honeywell- Brown recorder controller activated by the pressure drop across the orifice. The controller will control any flow from 40 to 300 gpm while the reading shown on the recorder is correct from 100 to 300 gpm. The orifice has been calibrated with a weigh tank from 5 45 to 200 gpm using a 1.750 specific gravity oil up to 50 inches oil (2.98 in Hg) and a mercury manometer from 2.5 to 6.0 inches of mercury (maximum flow sustainable with pumps available at that time). Since the calibration curve is straight from 0.2 in. of mercury up with a slope of 0.5, it has been extrapolated for the rest of the flow range. A steam heat exchanger is used to pre-heat the water. The steam flow is controlled by a valve activated by a Minneapolis-Honeywell-Brown recorder controller such that the temperature leaving the pre-heater is constant. Calibrated thermometers placed in water filled wells are used to measure the water temperature both into and out of the experimental apparatus from 46 ~F to 116 ~F. outside this range copper-constantan thermocouples are used. The temperatures below the tube bundle are measured with six chromelalumel thermocouples shielded from radiation from below. The temperature above the bundle is measured with five chromel-alumel thermocouples placed so that linear averaging of temperature will give the correct stack temperature. Samples of the stack gas are aspirated through a perforated length of one-half inch iron pipe spanning the stack at the centerline of the stack. Figure 4 shows the apparatus from the left side. The controller at the top of the left rack is the water inlet temperature controller recorder. Directly below this is the water inlet flow controller recorder. The recorder in the rack to the right of the one referred to above continuously records the temperature in the stack above the diffuser. Behind the controllers on the floor is the steam heated exchanger used to pre-heat the water from the cooling tower. 13

0 4 (') X k (I o — 14

On the cabinet top at the left end are the potentiometer and selector switches used to measure the thermocouple temperatures. To the right is the Orsat gas analysis equipment used for analyzing the flue gas. The Rockwell Model 4 gas meter sets on the floor behind the gas fired heat exchanger and can be seen in the right of Figure 1. The Metrol gas flow instrumentation panel can be seen at the extreme right of Figure 1. 15

OPERATION OF EQUIPMENT Before starting up the equipment, it was necessary to turn on the cooling water pump. Once this was turned on, the drain valves on the equipment were closed and the main water valves were opened. The water flow and water temperature regulators were set to the desired settings and then turned on. Then the stack temperature recorder was turned on. The steam valve for the water pre-heater was opened and the condensate bled from the steam pipe before the bleed valve was closed. Before firing the burner, the three permanent pilot lights were checked and lit if not on. Then the power to the safety system was turned on and 30 seconds allowed for the tube to warm up. When the tube warmed up, the relay would drop in and the indicating light would come on. At this point the gas was turned on. After the main burners came on, the gas differential regulator was set to the de — sired setting and set on automatic control. Once the burners were on, the stack cover was removed. (It cannot be removed earlier or the stack starts a heavy down draft which is very difficult to reverse.) If the desired water flow rate was below 45,000 lb. /hr. (160 gpm), the manifold valves for the manometer using 1.750 SG oil were opened. This manometer is necessary to give sufficient accuracy below 2.6 inches of mercury across the orifice. Sufficient time must be allowed for the system to reach equilibrium. The usual period was nearly one hour. During this time, the information on ).a-ta Sheet No. 1, shown in Table IV, was filled in. It should be noted that this information was assumed constant for the entire run (about four hours). Space is also provided so that these sheets can be partially prepared in advance to serve as a record of future runs, e.g., desired water and gas flow-rates. It should be noted that although there is a water temperature regulator, there is also need of manual intervention. The water temperature is regulated by control of the steam fed to a pre-heater. This pre-heater can raise the water temperature about 15 ~F at a flow rate of 150 gpm of water. The normal temperature desired is between 60 ~F and 75 ~F. Without the cooling tower turned on, the heater will raise the sump water temperature about 10'F/hr with maximum gas flow. With the cooling tower on in the winter, the equilibrium temperature of the sump water may go down to 45 ~F. Thus, to maintain an inlet temperature of, say 70'F, one must turn the cooling tower fan on and off at alternate hoursl relying on the regulator to eliminate the sinusoidal variation of inlet temperature. The data for the experimental run is collected on Data Sheet No. 1, Table IV, and Data Sheet No. 2, Table V. Because of the amount of data to be collected, two people are necessary, one to collect the gas samples and analyze them through 1 6

the Orsat, the other to collect the other readings. This normally requires 30 to 45 minutes per run. A new set of data is started every hour and three sets of data are usually required to prove steady state conditions. Tables IV through VII present a complete sample set of data for experimental run No. 7 on December 10, 1964. Note that although there is a column under stack analysis for infra-red analysis, it is not presently being used. The shut-down procedure is essentially the reverse of the start-up. The gas regulator is set to manual and the flame is reduced to minimum, then the power to the safety interlocks is shut off and the gas main closed. The regulators are turned to zero and shut off, then the water valves are closed. The system drain valves are opened and left open. Then the steam valve is closed, the water pump shut off and the stack cover replaced. During the run, samples of the stack and feed gas are taken for the mass spectrograph. A present, samples of the stack gas are taken twice during a set of runs and one sample of the feed gas is taken for the set of runs. Data Sheet No. 2, Table V, provides a place for recording the amount of condensation evident to the operators during the experimental run. 17

TABLE IV Hot Gas to Fin Tube Heat Transfer Data Sheet Noo 1 Run No. Date Operators 7 12/10/64 C. W. Bash Time 12:00 noon M. V. Kulkarni and Wet Bulb Temperature Dry Bulb Temperature Humidity 0 007 59.8 ~F 76.6 ~F lb. water/lb. dry air Barometric Pressure Temperature Correction Calibration Correction Corrected Pressure Water Thermometer No.: Water Thermometer No.: Water Orifice No. Used Desired water flow Desired gas flow 29.46 in -0.122 -0.006 29.332 in 89164 out 55079 #3 100 30. Hg at 74 ~F in. Hg at 32 ~F gpm = cfm in. Hg Regulator Settings Water flow rate 100 Water inlet temperature Gas flow differential gpm 72 ~F in. water 2.0 18

TABLE V Data Sheet No. 2 Run No. Date 7A 12/10/64 Time 2:30 AI/PM Flow Conditions: Water T from tower 70 ~F T into heater 68.6 ~F T from heater 97.5 ~F AP across orifice 2.2 in Hg 39.3 oil Gas T. in - P. in - regulated - AP (meter) 100 in 3 min 100 ft in 3 mmin 28.8 14.5 8.0 ~C in water in water 3.4 in water 13.55 sec 31.0 cfm Condensation none Stack Temperature light 190 heavy Thermocouples EMF 1 35.75 1 2 31.05 1 3 31.40 1 4 31.7 1, 5 31.7 17 6 31.55 1. T 580 374 389 402 402 396 (T. - T. ) 1 i-i EMF ~ — 7 5.79 -- 8 8.85 -- 9 6.87 -- 10 6.61 -- 11 5.34 -- -water in water out water temp rise = T 287 424 335 324 267 (T - T. ) i i-1 28.9 OF, Or sat Analysi IR s: Full_ co —-- CO2 02 2C 13.5 CO 13.5 Level 99.30 93.24 83.29 83.30 AL 6.06 9.95 Mass Spec. Sample 7-A H/C = 4.02 Excess Air = 82 % 19

TABLE VI Data Sheet No. 3 Run No. Date Time 7B 12/10/64 3:15 A4 PM Flow Conditions: Water T from tower T into heater T from heater AP across orifice 67. 96. i I - 65 ~F 8 OF 8 ~F 9 ~F 2 3 in Hg 39.6 in oil Gas T. in P. ] in regulatedAP (meter) 29.1 4.5 8 3.35 ~C in water in water in water 100 ft in 3 min 13.3 sec 31.0 cfm Condensation None Stack Temperature Thermocouple s EMF T 1 36.0 1591 2 31.3 1385 3 31.6 1398 4 31.9 1411 5 31.05 1374 6 31.8 1407 light 200 (Ti - T ) 1 1-1 +11 +11 + 9 + 9 -28 +11 heavy oF ) EMF T 7 5.86 290 8 8.9 426 9 6.91 337 10 6.75 330 11 5.33 266 water in water out water temp rise = 29.:1 (T,-T. ) 1 i. l +3 +2 +2 +6 -1 Orsat Analysis: IR Level 99.17 AL Full CO2 ~2 CO 93.12 6.05 83 043 9.70 24.5 83.41 0.02,~ ~ ~~~ ~ ~~~ ~ ~~~ ~ ~~~ ~ ~~~ ~ ~~~ ~ ~~~ ~ ~~~ ~ ~~~ ~ ~~~ ~ ~~~ ~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ H/C = 4. Excess Air = 26 77.5 ZQ

TABLE VII Data Sheet No. 4 Run No. Date Time 7C 12/10/64 4:15 A^ —/PM Flow Conditions: Water T from tower T into heater T from heater AP across orifice 62.2 67.5 96.3 2.4 40.2 ~F ~F ~F in Hg in oil Gas T. in P. in regulated AP (meter) t in 3 1Of t in 3 29.3 14.5 ~C in water in water in water 8.0 3.4 _ min 13.5 sec 31.0 cfm Condensation none Stack Temperature light 190 ~F heavy Thermocouple s EMF 1 35.75 2 31.10 3 31.45 4 31.70 5 30.90 6 31.45 T 1580 1366 1392 1402 1368 1392 (T. - T ) i i-1 -11 -19 - 6 - 9 - 6 -19 EMF 7 5.80 8 8.86 9 6.87 10 6.67 11 5.26 water in - water out -- thermocouple temp rise thermometer temp rise T 287 424 335 326 263 (T. -T ) 1 1-1 -3 -2 -2 -4 -3 28.8 Mass Spec. Samples 7 C Stack 7 Feed Stack Gas Analysis: IR Level Full -- 99.9 CO -- 93.9 2 0 -- 83.8 2 CO -- 8307 AL 6.00 10.10 0.10 H/C = 4.04 Excess Air = 82.5 21

STACK GAS AND FUEL GAS ANALYSIS Samples of the stack gas were collected at both ends of a perforated section of 1/2 in. pipe spanning the stack. The two lines were connected to an aspirator to continually purge the lines. Provision was made for collecting samples of stack gas for Orsat analysis and mass spectrometer analysis. Although a very old and often studied problem, analysis of stack gases is still not an exact science whenever the gas contains large amounts of water vapor. Since both analyses take place below the dew point of the flue gas, it is necessary to remove most of the water vapor before analysis is made. This introduces problems because of the solubility of carbon dioxide and carbon monoxide in the water removed. The first method of analysis attempted was the use of the basic Orsat apparatus for analysis of carbon dioxide, oxygen and carbon monoxide. The residue is assumed to be nitrogen. Knowing these four quantities and the composition of the feed gas, the ratio of volume of air to volume of natural gas at the burner can be calculated. Very poor reproducibility was obtained and it was impossible to get results that would correlate with the known feed gas composition as supplied by Michigan Consolidated Gas Company branch in Ann Arbor. It was next decided to use a mass spectrometer to determine the stack gas composition. At the same time, samples of the fuel gas would also be analyzed. Figure 5 shows the mass spectrometer apparatus used in making these analyses. The mass spectrometer takes a small sample of gas and breaks the molecules into smaller charged ions. These ions are passed through an electromagnetic field that curves the path of the ion in a manner proportional to the mass-to-charge ratio. At the end of the path a current measuring device counts the number of ions with a specific mass-to-charge ratio. By varying the electrical field, one can continuously scan a range of molecular weights. The amount of current measured is automatically plotted by a galvanometer. A typical plot of a fuel gas analysis is presented in Figure 6. It is known that the way a molecule breaks up is a characteristic of the molecule and independent of other species present. Also, the height of a given peak is directly proportional to the partial pressure of the compound in the sample. Since no two compounds break up in the same way, a set of simultaneous equations for the contribution of each component to the height of each peak can be solved for the partial pressure of each component present. This calculation is done on the IBM 7090 computer using a program written in the Michigan Algorithmic Decoder (MAD) language. At present, there is not sufficiently good agreement between the mass spectrometer results and the gas company method of analysis. The feed gas 22

q) 4: p4 w N ~ 04: 4) 4) 0 4-) 04 (n).r4

U) 0 H aQ CD J OQ.II 09 *ll 00 N 0 N c3 I. 0) P U) cn CD o ct 0 CD ct i-I 0 <Ch I ii ~r ~'r::1 I i i i j i i i i I i i: i ji j r i r i i ii i I i i i ~, 1 i i 1 i i I i i I I i i I I j i r I I'j i I i i? i i I i i i i: I i I i I i i 1 j i I 1 i i i j-l ~ ir ii I i i i I i I I I' i i i I r i i i i j i t i 1 i: i r i j j p I - ~':I I i I i i i I i. . I r i i j i~i i I I ( I 7 i i r i 1 i I I I 1 i ii j j ~ i i i: j' t i i: I i: i i i i i j: i ji; ~ I I: i. i i I i i... 1. I I i i Hi f i2. ii1 li I I i i ii " i l * I II I i l j i. 1 i i 1 1 1 s I'i i I....... "1. iI~ 1i.......,,,,, t i I I i ri r j z I i~ I i - j;/ r i i j i i I I' I j r:r t r r i i I i I r i i i 1 i I r;;i I i i i rp i r I I:.1..;I......:;.i.j i i I lil II -t 4 0 ( Pc O 4 U.) 4P

composition according to the mass spectrometer is always lower in methane and higher in propane than that indicated by the gas company. Table VIII presents a comparison of the mass spectrometer results with the composition provided by the gas company. TABLE VIII Comparison of Fuel Gas Compositions Component Gas Co. Mass Spectrometer Run Number 14060 14102 14101 Methane 93.50 Ethane 3.85 87.18 6.92 1.34.17 84.71 86.41 8.78 7.75 Propane i- Butane.77 1.80 1.80.08.16.12 n- Butane Nitrogen Carbon Dioxide Others Gross Heating Value Net Heating Value Mole of Carbon/Mole of Fuel Mole of Hydrogen/Mole of Fuel H/C, atomic ratio.01 0.05.04.36 4.38 4.24 3.71 1.31 0.26.36.12 0 0 0 1033 1099.4 993.3 1069.1 1058.4 941 966.4 956.8 1.0518 1.0572 4.0416 4.0266 1.0851 4.0802 3.760 1.0774 4.0660 3.774 3.843 3.809 It has also been found that the mass spectrometer results do not agree exactly with the Orsat analysis of the flue gas, but the discrepancy is within reasonable error. A comparison of two typical analyses is shown in Table IX. An explanation of how the mass spectrometer results are calculated is presented in the Appendix. 25

Carbon Dioxide Oxygen Carbon Monoxide Nitrogen & Argon H/C % Excess Air TABLE IX Comparison of Flue Gas Analyses Run 7 C Orsat Mass Spec. 6.0 6.08 10.1 9.79 0 0 83.7 83.08 4.04 4.097 82.5 79.59 Run 9 B Orsat Mass Spec. 7.89 7.34 6.90 7.25 0 0 84.80 85.37 3.92 4.27 44 3 47.96 26

ANALYSIS OF DATA Heat Balance The performance of the water heater is evaluated on the basis of a heat balance. The calculation involves the following: 1) The total heat input, calculated from the fuel gas input flow rate and its heating value. 2) The total heat transfer to water, calculated from the total flow rate and rise in temperature of the water. Care was taken to calibrate the water orifice and the thermometers to minimize the source of experimental error. 3) The convective heat transfer. (The total heat transfer is the sum of radiative heat transfer and convective heat transfer.) The convective heat transfer is calculated from the temperature drop of stack gas across the water tube bundle as measured by thermocouples and the gas flow rate computed from the stack gas analysis and fuel gas rate. 4) Radiative heat transfer, the difference between the total heat transfer and convective heat transfer is taken as being due to radiation. 5) Total heat loss and efficiency of the water heater is then calculated by usual procedures. Radiation The radiative heat transfer for a typical analysis amounts to about 35 to 45 percent. A survey of the recent literature published on the design and performance of fired heaters (11, 18) indicates that the radiation heat transfer should be less than this amount. The basis for radiant heat transfer is the Stefan-Boltzman equation. A black body at absolute temperature T radiates energy at a rate of qb by the following relationship: q,= o T4 (1) 27

and for radiant heat transfer between two real surfaces at a temperature T and Tb, thie relation becomes: a 4 4 q= A (F T - 4) (2) Ar Fab a a b b where ~- ~ = 1.712 x 10 9 Btu/hr.ft ~R A = Area of surface a, sq. ft. F a= Interchange shape factor from a to b ab E, e = Emissivities of surfaces a and b a b T, T = Absolute temperatures of surfaces a and b, ~R a b The luminous radiant heat transfer rate calculated from the above equation using different shape factors for our particular water heater is given in Figure 7. A small amount of heat transfer also takes place due to non-luminous radiation from the gas in the convection zone following the method presented by Hottel (36). For finned tubes the gas radiation coefficient is based on root diameter of the tube. The area used is the area of bare tube. A tube surface emissivity of 0.9 and a constant partial pressure of CO and H20 and a tube-to-tube distance of 1.5 to 2 diameters were assumed. The apparent heat transfer coefficient for non-luminous radiation from the flue gas to the tube is given by Equation 3. (1 + et) 4 (".e) 4 4 (h) 2 (e T - e T ) h = tb............ (3) (T- Tt) g t where the average gas temperature T is defined as the average inlet gas temperature plus the log mean temperature difference from gas to inlet water, the average tube wall temperature Tt is approximately the average inlet water temperature plus 50 ~F and: e emissivity of tube = 0.9 t e emissivity of the flue gas at the temperature in g question at a partial pressure of (water + CO2) of.19 atm (18) 28

1.0 0.9 0.8 0.6 0.5 0.4 0.3 0.2 0.15 0.1 0.09 0.08 0.07 1400 1600 1800 2000 2200 2400 2600 Flame Temperature, ~F Rate of Luminous Radiation Heat Transfer Fig. 7 29

The solution of Equation 3 for 80% excess air is presented in Figure 8. This figure indicates that non-luminous gas radiation should not be neglected. Example Analysis The summary of a typical analysis (data collected for Experimental Run No. 7) is presented in Table X. A review of the results presented in Table X indicates that the radiative heat transfer is excessively high compared to theoretical predictions. This raises the question as to whether or not the thermocouple temperature measurements of the hot flue gasses are correct. Subsequently, a theoretical analysis of the radiation losses from the thermocouples to the cold tubes was made. This analysis showed that the indicated temperatures below the bundle were low by approximately 200 ~F. On the basis of these theoretical calculations, the heat balances were revised by raising the measured hot flue gas temperatures as measured by thermocouples located below the tube bundle by 200 ~F. TABLE X Analysis of Data Collected for Experimental Run No. 7 Run 7 A B C Fuel gas flow rate S. cu.ft./mi. 30.24 30.20 30.06 4 Water flow rate, 10 lbs./hr. 4.85 4.85 4.8 Stack gas flow rate, lbs./hr. 2631 2658 2620 Water temperature rise, ~F 28.9 29.1 28.8 Stack gas temperatures (Ave.), ~F Below the water tube bundle 1424 1428 1420 Above the water tube bundle 327.4 329.8 327 -6 Total heat input, 10 x Btu/hr. 1.71 1.702 1.698 Total heat transfer to water, 10 x Btu/hr. 1.401 1.410 1.382 Convective heat transfer, 10 x Btu/hr. 0.842 0.824 0.802 Radiative heat transfer (by difference) 10-6 x Btu/hr. 0.559 0.586 0.581 -6 Total heat lost, 10 x Btu/hr. 0.309 0.29 0.316 -6 Sensible heat lost, 10 x Btu/hr. 0.168 0.173 0.168 -6 Radiation heat lost, 10 x Btu/hr. 0.141 0.117 0.148 Efficiency, per cent 82.1 83.1 81.51 Heat Transfer by Radiation, % 39.7 41.8 41.8 Excess air, % 82 77.5 82.5 30

1.6 1.4 Z 1.2 0,<a I~~.Z Tube Wall Temperature = 220 ~F C79^1.0 kTube Wall Temperature = 150 ~F 0.8 - U o 0.6ro rI 0.4 600 700 800 900 1000 1100 1200 Average Gas Temperature, ~F Fig. 8 Non-luminous Gas Radiation Heat Transfer Coefficient 31

The revised heat balance on the basis of the estimated hot gas temperatures for Run No. 7 are presented in Table XI. TABLE XI Revised Heat Balance for Run No. 7 A B C Total heat input, 106 x Btu/hr. 1.71 1.702 1.698 Total heat transfer, 10 x Btu/hr. 1.401 1.410 1.382 -6 Convection heat transfer, 10 x Btu/hr.,Q 0.976 0.983 0.972 -6 CRadiation heat transfer, 10 x Btu/hr.,Q 0.425 0.427 0.410 r Per cent heat transfer by convection 69.6 69.7 70.3 Per cent heat transfer by radiation 30.4 30.3 29.7 A comparison of the results given in Table XI with Table X indicates that the radiative heat transfer has been reduced by approximately one-third. The overall heat transfer coefficient is given by Equation 4. Q = U A At (4) c o o m or re-arranging: Q U = c (5) o A At o m where: Q = convective heat transfer, Btu/hr c 2 U = overall heat transfer coefficient, Btu/hr.-ft. F o A = heat transfer area of fin tube, sq.ft. o At = log mean temperature difference, ~F m 32

A summary of overall heat transfer coefficients in Run No. 7 are as follows: for the revised data taken Q, Btu/hr. c A, sq.ft. At, ~F m U, Btu/hr-sqft- ~F 0 A 0.976 x 106 144.75 714 9.45 B 0.983 x 106 144.75 717 9.48 C 0.972 x 106 144.75 715 9.40 The individual gas film heat transfer coefficient can be calculated after determining the individual resistances that make up the total resistance, 1/U U Water Film Coefficient The water film coefficient is given by (36): 0.8 (Vt) h. = 150 (1 +.011 t )! w (Dw 0.2 (D) i (6) where: h. 1 t w V t D. 1 = water side film coefficient, Btu/hr-sqft- ~F =mean water temperature, ~F = velocity of water inside the tube, ft/sec = inside diameter of tube, inches A summary of h. for Run No, 7 is as follows: 1 mean water temperature, t, ~F w water flow rate, cu ft/sec area of flow, sq ft Vt, ft/sec D., inches 1 ho, Btu/hr-sq ft- ~F 1 A 83.05 0.216 0.0125 17.3.875 2,890 B 82.35 0.216 0.0125 17.3.875 2,880 C 81.9 0.214 0.0125 17.1.875 2,840 33

Gas Film Coefficient The gas film coefficient is determined from Equation 7. - = -h + rf + rw ( A+ hT (A) 1 1 A 1 o0 m where A = mean area in sq. ft. /ft. m U = overall heat transfer coefficient, Btu/hr-sq ft- ~F o h = actual gas film coefficient, Btu/hr-sq ft- ~F o r = resistance of fin r = resistance of root wall w h. = water side film coefficient (Equation 6) 1 A A = ratio of outside surface to inside surface A. 1 The fin resistance is given by Equation 8. (7) rf 1 m /3 D /D m + 3 Dr [ m 2 ~7m2 /D~ K'^vr) (8) where m = H 2 (9) 1 (km) (y) h + ro 0 34

Assuming different h' values, rf, the fin resistance can be calculated for the particular fin. The fin resistance calculates out to be a constant value of 0.00025 for h' of 1L0 to 100. o The metal wall resistance is given by: AA rn X * A X A k *A m where k = thermal conductivity = 196 Btu/hr-sq ft- ~F m A = 2.316 sq ft/ft O D = 1.0 inch r D. =.875 inch 1 X = thickness of wall =.063 inch D - D. 1 -.875 \ A = r r r-1 () - -.875 =.2455 sq ft/ft m n (D /D r 1.875 A, I9.063 ) i2.316\ 196 x 1 2/.2455.0002525 The calculated values of h' for the revised data for Run 7 using Equation 7 are presented in Table XII. 35

U 0 o rf A w A ) m / A. 1i( A ) h' o TABLE XII Calculated Coefficients and Resistances A B 9.45 9,48 0.00255 0.00255 0.00025 0.00025 C 9.40 0.00255 0.00025 281 10.0 286 10.0 285 10.15 36

DISCUSSION OF RESULTS Considerable difficulties were encountered in obtaining reliable physical property data for the components of the flue gas in the temperature range of 500 ~F to 2000 ~F. Figures 13 through 24 present the physical property data as they are presently being used by the project staff. The thermal conductivity of carbon dioxide has been extrapolated. In extrapolation, an attempt was made to make the curve parallel to Geringer's data (35). Geringer's data appears to be consistently lower than the NBS data (31). The thermal conductivity of water vapor has been extrapolated using the equation of Keenan and Keyes (34). Thermal conductivity data for oxygen apparently is available only to 1000 ~F. However, since the available data is nearly a straight line, it has been extrapolated to 2000 ~F. The estimation of the physical properties of a mixture of gases is somewhat of a problem. There is a dispute in the technical literature as to the proper relationships that should be used. Heat capacity of the flue gas was assumed to be a linear function of the weight fraction and was calculated using Equation 10. C =C WC W +C W...C W (10) Pm P1 1 P2 2 n where C is heat capacity in Btu/lb- ~F P W is the weight fraction subscript m implies mixture subscripts 1 through n are components 1 through n Viscosity of the flue gas is estimated by Equation 11 according to Maxwell (37) and Perry (38). n - I i.x i 1 ^ p-. x. M, i= 1...- (11) nr i 7xi I~~~~ See Appendix, pages 56-67. 37

where 1 X. 1 M, 1 = the viscosity in lb/ft hr for the i th component = the mole fraction of component i = the molecular weight of component i i = 1,2,..., n numbers identifying the components subscript m implies a mixture The prediction of thermal conductivity of a gas mixture requires the use of an extensive set of equations. These are reported by Reid and Sherwood (39) and for the sake of brevity are not reported here. Table XIII summarizes the various references used for determination of the physical properties of the flue gas components. TABLE XIII References Used for Physical Properties of Flue Gas Flue Gas Component Heat Capacity (Ref. No.) Viscosity (Ref. No.) N2 Thermal Conductivity (Ref. No.) 31, 33 31, 33 31, 32 31 02 31, 32 31 CO2 31, 32 31 31, 33, 35 HO2 2 31, 32 31 31, 33, 34 Although Table XIII summarizes the sources of the physical property information used for preparation of the physical property curves, some question still exists as to whether all of this information is reliable enough for analytical purposes for all components in this laboratory investigation. Theoretical calculations of the temperature of the flue gas leaving the burner as a function of the per cent excess air were made. Figure 9 presents the curve. Examination of this figure indicates that with 20% excess air the theoretical gas temperature could be expected to be 2780 ~F. With 80% excess air the temperature would be 2080 ~F. Table X indicates that the measured hot gas temperature for Run No. 7 with 80% excess air was 1425 ~F. The calculated 38

3600 3200 2800 0 X 2400 S \ H 2000 C w 1600 1-200 800 0 100 200 Percent Excess Air Fig. 9 Theoretical Temperature of Gas Leaving Burners 300 39

theoretical gas temperature assumes 100% combustion efficiency and no radiation affects. The calculations for Run No. 7 indicate that if the 82% efficiency and the direct radiation encountered are taken into account, then the measured gas temperature should have been approximately 1600 ~F or higher. The dew point of the hot flue gases is a factor in the performance of the experimental unit. During all experimental test runs, except for one, condensation of water on the finned tubes was encountered. The condensation affects the heat transfer performance. Theoretical calculations of the dew point of the flue gas assuming no water vapor in the combustion inlet air were made. Figure 10 presents the dew-point curve. The curve indicates a dew point of 132 ~F for 20%o excess air and a dew point of 119 ~F for 80% excess air. This problem is aggravated on humid days. Condensation increases the heat transfer rate but corrodes the finned tubes. As noted earlier, no general correlation for predicting the heat transfer coefficient for banks of finned tubes on a rectangular pitch exists. For comparison purposes, the correlation of Briggs and Young (15) for equilateral triangular pitch banks of finned tubes was used. The theoretical hot gas temperature was read from Figure 9 with an average water temperature of 100 ~F assumed and gas film coefficients calculated. Figure 11 presents the predicted heat transfer coefficients as a function of natural gas feed rate with 20%o and 80% excess air. Table X indicates that the fuel gas rate was 30 cubic feet per minute for Run No. 7. Figure 11 would predict a coefficient of 7.4 for 80% excess air for triangular pitch banks of this finned tube with the same pitch. Table XII indicates that the calculated gas film heat transfer coefficient from the experimental data for this run was 10. This value is 35% higher than that predicted by the triangular pitch correlation. The actual experimental gas film coefficient for exper.imental Run No. 7 should be less than 7 4. The experimental laboratory furnace was also instrumented for measuring the pressure drop of the flue gas flowing across the bank of finned tubes. To date no satisfactory reproducible measurements have been made. It is thought that the great change in density of the hot flue gas as it cools down in flowing across the finned tubes is upsetting the measurements. Various locations of the presure taps are being tried out in an attempt to obtain more realistic and reproducible results. A micro-manometer is being installed in an attempt to obtain more accurate results. However, it is believed that the fluctuations in gas pressure may possibly rule out the use of a micro-manometer. The experimental data obtained to date indicate that the manufacturer of the experimental unit is correct in claiming that the unit operates at 80% efficiency. 40

150 140 130 - 120 r-4 ~o 110 100 - 90 80 0 100 200 300 Percent Excess Air Fig. 10 Dewpoint of Flue Gas 41

9 8 7 5 lCI 4- a 4 3 Gas Feed Rate, cu.ft /min. Fig. 11 Gas Film Coefficient for Triangular Pitch Tubes 42

In summary, the following situations were causing some difficulties at the time of the preparation of this progress report: (1) Disagreement with the gas company on the actual composition of the natural gas used as fuel. (2) The effect of the above disagreement on the analysis of the experimental data. (3) Some disagreement, but not serious, between the Orsat analysis and mass spectrometer analysis of the "stack" or "flue" gas. (4) Difficulty in accurately measuring the temperature of the hot flue gas (burner gas) below the finned tube bundle without radiation affects from the burner flames and re-radiation from the walls. (5) The analysis of the experimental data obtained to date indicates that the radiation heat transfer is in excess of that predicted from theory. (6) The analysis of the experimental data obtained to date also indicate that the calculated convective heat transfer coefficients may be too high. (7) Some question exists as to whether the physical property data currently being used for some of the flue gas components are reliable at the elevated temperatures. (Some of the data are calculated from molecular model theory, others are semi-empirical.) (8) Condensation of water vapor on fins below the dew point of the flue gas is upsetting the heat transfer since it is impossible to accurately collect and weigh the condensate formed. (9) The per cent excess air measured in the experiments to date are believed to be considerably higher than those normally encountered in hot water heater installations. (10) Difficulty is being encountered in obtaining satisfactory pressure drop measurements of the hot flue gas across the finned tubes. The above ten items were being worked on at the time of the preparation of this progress report. 43

MODIFICATIONS TO THE EXPERIMENTAL APPARATUS AND FUTURE WORK The use of multiple metal radiation shields around the thermocouples should reduce the radiation error by reducing the net rate of radiation from the thermocouple. Figure 12 shows a sketch of the radiation shields that will be used. These shields are patterned after a design suggested by Moffatt (40) and King (41). In order to separate the heat transfer into the effect of each row, it was necessary to redesign the water flow from a four-pass design to a one-pass design. This necessitated fabrication of new water headers and tube sheets. Table XIV shows a list of the tubes that are on order from Wolverine Tube for experimental investigation. Performance of these tubes will be investigated in both triangular and square pitch arrangement until sufficient experimental data is available to rule out one of these pitch arrangements. The effect of the number of horizontal rows will be investigated by collecting experimental data on each horizontal row, then cutting out one horizontal row of tubes at a time and plugging the tube holes in the tube sheets with brass plugs. TABLE XIV Tubes on Order for Investigation Tube No. Rows Tubes/Row Bundle No. 61-0714065 2 6/pass (4 passes) 1 61-0714065 3 13 2 61-0710065 4 16 3 61-0720065 not determined 11 61-0516065 3 14 4 61-0916065 (if available) 3 13 7 or 61-0920065J 61-0520065 3 11 5 61-0714072 3 13 6 61-0910049 4 16 8 Largest available ID and fin height 61-0905035 4 26 10 Thermocouples will be inserted through the exit water header cover plate into the tubes so that the temperature rise for the water in each row of tubes can be measured. 44

B.S. 11 gage Chromel Alumel Thermocouple.5I I I I in In N o ~ 2 JLfL)o O o un o uLn N dso ro ) -4 ~ ln ~ L- ~2.50"' - J Fig. 12 Sketch of Thermocouple Shield Assembly 45.

LITERATURE CITED 1. Griffiths, J. C. and Pountney, C. H., Jr., "The Application of Heat to Domestic Gas Storage Water Heaters," American Gas Association Laboratories Research Bulletin 71, Cleveland, 1956. 2. DeWerth, D. W. and Smith, Ro E., "Design Factors of Gas Appliances for More Effective Use of Heat Exchanger Surface," American Gas Association Laboratories Research Bulletin 86, Cleveland, 1961. 3. American Gas Association Laboratories, "Effect of Cold Inlet Water on Performance of Automatic Storage Gas Water Heaters," Research Bulletin 19, Cleveland, 1943. 4. American Gas Association Laboratories, "Principles of Design and Sizing of Automatic Gas Water Heaters for Maximum Service Efficiency," Research Bulletin 29, Cleveland, 1944. 5. American Gas Association Testing Laboratories, "Research in Special Types of Burners for Flames Adaptable to Gas Water Heaters," Research Bulletin 45, Cleveland, 1947. 6. Griffiths, J. C. and Weber, E. J., "The Design and Application of Impingement Target Burners," American Gas Association Laboratories Bulletin 75, Cleveland, 1957. 7. Hammaker, F. G., "Gas Company Experiences with the Experimental High Performance Burner," American Gas Association Laboratories Bulletin 80, Cleveland, 1960. 8. American Gas Association Laboratories, "Fundamentals of Heat Transfer in Domestic Gas Furnaces," Research Bulletin 63, Cleveland, 1951. 9. American Gas Association Laboratories, "Study of Performance Characteristics of Gas Boilers Equipped for Hot Water and Space Heating Services," Research Report 1185, Cleveland, 1952. L0. American Gas Association Laboratories, "Condensation in Gas Water Heaters," Research Bulletin 58, Cleveland, 1950..1. Schweppe, Jo C. and Torrijos, Co Q., "How to Rate Finned-Tube Convection Fired Heaters," Hydrocarbon Processing & Petroleum Refiner, 43, No. 6, 159-166 (1964). 4-4

12. Torrijos, C. Q. M.S. Thesis in Mechanical Engineering, The University of Houston (1963). 13. Lobo, W. E. and Evans, J. E., "Heat Transfer in Radiant Section of Petroleum Heaters, " Trans. Am. Inst. Chem. Engrs., 35, 743-778 (1939). 14. Monrad, C. C., "Heat Transmission in Convection Section of Pipe Stills," Ind. Eng. Chem., 24, 505 (1932). 15. Briggs, Do E. and Young, E. H., "Convection Heat Transfer and Pressure Drop of Air Flowing Across Triangular Pitch Banks of Finned Tubes", AIChE Chem. Eng. Prog. Symposium Series No. 41, Vol. 59, 1-10 (1963). 16. Gardner, K. A., "Efficiency of Extended Surface, " Trans. Am Soc. Mech. Engrs., 67, 621 (1945). 17. Gunter, A. Y. and Shaw, W. A., "A General Correlation of Friction Factors for Various Types of Surfaces in Cross-Flow," Trans. Am. Soc. Mech. Engrs., 67, 643 (1945). 18. Wimpress, R. N., "Rating Fired Heaters", Hydrocarbon Processing & Petroleum Refiner, 42, No. 10, 115-126 (1963). 19. Wilson, D. W., Lobo, W. E. and Hottel, H. O., "Heat Transmission in Radiant Sections of Tube Stills," Ind. Eng. Chem., 24, 486-93 (1932). 20. Mieth, H. C., "The Importance of Fintube Application in Energy Conservation," Paper 64-Pet-4, Petroleum Mechanical Engineering Conference, Los Angeles, California, Sept. 20-23, 1964, ASME. 21. American Standards Association, "American Standard Approval Requirements for Gas Water Heaters, " v. 1, American Gas Association, Cleveland (1963). 22. American Standards Association, "American Standard Approval Requirements for Gas Water Heaters, " v. 2, American Gas Association, Cleveland (1963). 23. American Standards Association, "American Standard Approval Requirements for Gas Water Heaters," v. 3, American Gas Association, Cleveland (1963). 24. American Standards Association, "American Standard Approval Requirements for Central Heating Gas Appliances," v. 1, American Gas Association, Cleveland (1962). 47

25. Kirk, W. Bo, "Flue Gas Temperature Measurement Procedures," Amer. Soc. Heat. Ref. and Air Cond. Engrs. Journal, 6, No. 6, 40-44 (1964). 26. Schlentner, K. O., "Flue Gas Temperature Measurement Procedures Followed by I.B.R. and S.B.I.," paper presented before The Flue Gas Temp. Meas. Symp., ASHRAE Semiannual Meeting, New Orleans, La., Jan. 27-29, 1964. 27. Barrett, Richard E. and Hazard, Herbert R., "Problems in Flue-Gas Temperature Measurements," Amer. Soc. Heat. Ref. and Air Cond. Engrs. Journal, 7, No. 1, 88-93, 104 (1965). 28. Ward, D. J. and Young, E. H., "Heat Transfer and Pressure Drop of Air in Forced Convection Across Triangular Pitch of Finned Tubes, " Chemical Engineering Progress Symposium Series No. 29, 55, 37 (1959). 29. deLorenzi, Otto, "Combustion Engineering," Combustion EngineeringSuperheater Inc., New York (1951). 30. North American Manufacturing Co., "North American Combustion Handbook," Cleveland (1952). 31. "Tables of Thermal Properties of Gases," comprising Tables of Thermodynamic and Transport Properties of Air, Argon, Carbon Dioxide, Carbon Monoxide, Hydrogen, Nitrogen, Oxygen, and Steam, by J. Hilsenrath et al. National Bureau of Standards Circular 564, 1955, UoS. Government Printing Office, Washington, D. C. 32. McBride, B. J., et al, "Thermodynamic Properties to 6000 ~K for 210 Substances Involving The First 18 Elements," NASA SP-3001 (1963). 33. Timrot, D. L. and Vargaftig, N. B., "Heat Conductivity, Viscosity, and Thermodynamical Properties of Steam at High Temperatures and Pressures," Transactions of Fourth World Power Conference, III, Lund-Humphries, London, 1649 (1952). 34. Keenan, Jo H. and Keyes, F. G.,' Thermodynamic Properties of Steam, John Wiley and Sons Inc., New York (1961). 35. Geiringer, P. L., Handbook of Heat Transfer Media, Reinhold Publishing Corp., New York (1962). 36. McAdams, Wo H., Heat Transmission, McGraw Hill Book Company, Inc., New York (1954). 48

37. Maxwell, J. B., Data Book on Hydrocarbons, D. Van Nostrand Co., Inc., New York (1950). 38. Perry, R. H., Chemical Engineers' Handbook, McGraw-Hill, New York (1963). 39. Reid, R. C. and Sherwood, P. K., Properties of Gases and Liquids, McGraw Hill, New York (1958). 40. Moffatt, E. M., "Multiple-Shielded High-Temperature Probes," SAE Quarterly Transactions, Vol. 6, No. 4, 567-580 (1952). 41. King, W;': J., "Measurement of High Temperatures in High-Velocity Gas Streams, " Trans. ASME, 65, 421-431 (1943). 42. Lapidus, L. W., Digital Computations for Chemical Engineers, McGraw Hill, New York (1962). 49

APPENDIX 5.Q

CALCULATION OF MASS SPECTROMETER RESULTS The mass spectrometer always breaks a molecule up in a manner that is distinct for a given molecule and the peak height is always proportional to the partial pressure of the compound in the sample. The usual method of analysis is to define a sensitivity for a given compound as the number of divisions (Figure 6) produced per micron of partial pressure at a particular mass-to-charge ratio (M/E). All other peaks for that compound will be indicated as the ratio of the height of that peak to the height of the standard peak. With these definitions, one can readily show that Equation 12 describes the height of an unknown peak. SXR +SXR +......SXR. 11 1 j 2 2 2j n n j = P j (12) where S. = sensitivity of component i 1 X. = partial pressure of component i 1 P. = peak number j in sample 3 R. ratio of the peak of compound i at (M/E) number j to the standard peak of compound i i = 1,2,...., n j = 1 2,...., n In order to solve for the value of Xo, we need n where n is the number of components in the mixture. independent equations These n equations are as shown in Equations 13a to 13n. slX1Rll + S2X2R21 +... SXR n n nl = P 1 (13a) SXX R +... SXR 1 1 12 2222 n n SXR + S XR +... SXR 1 1 ln 2 2 2n nn nn = P 2 (13b) = P n (13n) 51

These equations are summarized in matrix form by Equation 14. (R) (S) (X) = (P) (14) where: 11 21 f 0 Rnl 12 22 (R) (15) R R... R in 2n nn (S) (X)= (P) = (S, S,..., S ) (1 2 n (X, X..., X ) 1 2 n (16) (17) (18) (P1' P2''''' Pn) The sensitivity vector, (S)', and the partial pressure vector, (P), may be combined to define a new variable (Y) as shown in Equation 19. (Y) (S) (X) = (S X SX,Z... S X) 11 22' n n (19) Substituting Equation 19 into Equations 13a through 13n gives Equations 20a through 20n^ RllY + R Y + 11! 21 2... R Y nl n = p.t (20a) R Y R Y +... R Y 12 1 22 2n2 n R Y + R +... R Y In 1 2n 2 nn n = P 2 (20b) = P n (20n) 52

This set of equations can be recognized as a simple set of linear algebraic equations. Equations 20a through 20n are summarized in matrix notation in Equation 21. (R) (Y) = (P) (21) The theory of matrices (42) allows these equations to be solved very quickly on a digital computer as a large number of the non-diagonal elements of the matrix (R) are zero. A computer program has been written in the MAD language to solve these equations for any set of unknowns consisting of less than ten compounds. Provision is included to calculate the (R) matrix from a set of standards and then to use these standards to solve for the partial pressure and mole fraction of the compounds present in a set of unknowns. Tables XV, XVI, and XVII show a sample computer output for Runs 7A and 7C. TABLE XV Typical Digital Computer Output for Mass Spectrometer Analysis of Natural Gas Sample No. 14101 for Run 7, Feed on December 10, 1964 M/E Compound Sensitivity 89.0359 Peak Height 4900.00 Partial Pressure 54.783 Percent Composition 86.409 15. 0 Methane 28.0 Nitrogen 30.0 Ethane 44.0 Carbon Dioxide 132.8002 30.8868 159.1601 1278.00 155.70 83.20 2.353 4.915.229 3.712 7.753.362 29.0 Propane 43.0 n-Butane 42.0 i-Butane 163.3267 207.2247 77.1060 312.00 64.20 17.30 1.012 1.597.027.043.079.124 63.399 100.000 Sample Pressure Heating Value is 1054.711 Btu/cu.ft. at 60 ~F.000 53

TABLE XVI Typical Digital Computer Output for Mass Spectrometer Analysis of Stack Gas for Run No. 7A Sample No, 14106 for Run 7A, on December 1, 1964 M/E Compound 28.0 14.0 32.0 44.0 18.0 40.0 Monoxide, Carbon Nitrogen Oxygen Dioxide, Carbon Water Vapor Argon Sensitivity 128.1727 19.6831 98.7805 144.2210 93. 2000 163.8000 Peak Height 4830.00 806.00 499.00 398.00 2.30 81.80 Partial Pre ssure.000 40.949 5.050 2.760.025.499 49.283 49.230 Percent Composition.000 ": 83.089 10.248 5.600.050 1.013 100.000 Sample Pressure * Negative answers are considered zero Combustion Analysis H/C = 4.457 Excess Air = 86.560 54

TABLE XVII Typical Digital Computer Output for Mass Spectrometer Analysis of Stack Gas for Run No. 7C Sample No. 14107 for Run 7C, on December 1, 1964 M/E Compound 28.0 Monoxide, Carbon 14.0 Nitrogen 32.0 Oxygen 44.0 Dioxide, Carbon 18.0 Water Vapor 40.0 Argon Sensitivity 128.1727 19.6831 98.7805 144.2210 93.2000 163.8000 Peak Height 4780.00 810.00 479.00 434.00 2.00 82.00 Partial Pressure.000 41.152 4.848 3.009.021.501 49.531 49.230 Percent Composition.000 * 83.083 9.788 6.075.043 1.011 100.000 Sample Pressure - Negative answers are considered zero Combustion Analysis H/C = 4.097 Excess Air = 79.588 55

.050.046. 042 / CO p.038 4> U -.034 0 r —I E.030.026.022 600 800 1000 1200 1400 1\ Temperature, ~F Fig. 13 Thermal Condiuctivity of Carbon Dioxide 56

0.310 0.300 - 0.290 - / O 0.280 - / U c. 0270 0.260 0.Z50 600 800 1000 1200 1400 1600 1800 Fig. 14 Heat Capacity of Carbon Dioxide 57

.116.108 o100.092 ^.084 Cl).1 4-j 0 U.076-.068.060 600 800 1000 1200 1400 1600 1800 Temperature, ~F Fig. 15 Viscosity of Carbon Dioxide

.044.042.040 /.038 cn.034.4o.032.030.028.026 600 800 1000 1200 1400 1600 1800 Temperature F Fig. 16 Thermal Conductivity of Nitrogen 59

.290.285.280 oZ 0'.275/./ 4-I >( U o..270 e/ U r(.265.260.255 600 800 1000 1200 1400 1600 1800 Temperature OF Fig. 17 Heat Capacity of Nitrogen 69

.115.105..095 4.a (08 n.085 /.Oa U C().075.0 6 5 lIlllllllllll 600 800 1000 1200 1400 1600 ~800 Temperature, ~F Fig. 18 Viscosity of Nitrogen 61

.054 050 0 4'-H cn..042 4 — rt 0380 0 U.034.030,026 600 800 1000 1200 1400 1600 1800 Temperature, ~F Fig. 19 Thermal Conductivity of Oxygen 62

r //.0250 P-1'.0245&.0240.0235 600 800 1000 1200 1400 1600 1800 Temperature, ~F Fig. 2D Heat Capacity of Oxygen 63

136.128.120. 112.104 - / 0 O.096.088.080 600 800 1000 1200 1400 1600 1800 Temperature, ~F Fig. 21 Viscosity of Oxygen 64

.060 I I I I.056-.052.048 4 4 —1 <4'.044.04 - 4CO 0.036 o.,c.032 -.028.024 600 800 1000 1200 1400 1600 1800 Temperature, ~F Fig. 22 Thermal Conductivity of Water Vapor 65

0.600 0.580 0.560 Po, 0.540 -- 0.520 0.500 0.480 600 800 1000 1200 1400 1600 Temperature, ~F Fig. 23 Heat Capacity of Water Vapor 66 1800

.112.104 -.096 -.088 o.080.064.056.048 / I I I I I I I 600 800 1000 1200 1400 1600 1800 Temperature, ~F Fig. 24 Viscosity of Water Vapor 67